Comparative analysis of hybrid and metaheuristic parameter estimation methods of solar PV

Abstract

The precise modelling of solar photovoltaic (PV) cell is very essential before proceeding to the installation of an actual solar PV system. This aids in simulation and emulation of a solar PV system for actual operating conditions and thus helps in enhancing the overall efficiency of a PV system. However, the exact knowledge of undermined solar PV cell’s parameters, i.e., photon current (Ipv),diode dark saturation current(Is),series resistance(Rs),shunt resistance(Rsh),and diode ideality factor(A) are highly prerequisite to obtain the accurate mathematical modelling. In this paper, three different manufacturer’s datasheet-based parameter estimation methods, namely the metaheuristic parameter estimation method i.e., practical swarm optimization (PSO), and two existing hybrid parameter estimation methods have been discussed and examined to reveal their effectiveness and accuracy. The hybrid methods used in this study optimises the parameterised model of single diode model (SDM), which is a function of only two undetermined parameters Rs and A. The Newton-Raphson (NR) and simulating annealing (SA) algorithm are used by the two hybrid methods under test, respectively. Further, to check the efficacy of the tested methods, two different technologies based solar PV modules, namely STM 40–36 (mono-crystalline) and STP6 120–36 (poly-crystalline), have been considered. Using these techniques, firstly, undermined model parameters are estimated and then output characteristics, in terms of I-V and P-V curves, are evaluated to obtain benchmarks, namely relative maximum power error (EARMP%), and overall model error. Lastly, by means of these benchmarks, it is found that the hybrid method which uses SA for estimating Rs and A, outperforms the other two approaches. The benchmarks, obtained with the SA based hybrid method, are smaller than the other two methods. Overall, it is observed that hybrid methods can be considered as more accurate parameter estimation technique and it can be considered to achieved more accurate mathematical modelling.